1、 2016 ASHRAE 89ABSTRACTCommercialandindustrialrefrigerationsystemsconsumeasignificant portion of U.S. electrical energy. In this research,advanced expansion valve and control algorithms are evaluatedto quantify the potential energy savings due to improved systemregulation and efficient start-up of v
2、apor-compression refriger-ation systems. The performance of the new micro-electro-mechanical system (MEMS) actuators with different controlstrategies is compared with that of the standard mechanicalvalvesandacommerciallyavailablesuperheatcontroller.Addi-tionally,thisresearchincludesacomprehensiveset
3、ofexperimen-talteststhatidentifythemosteffectiveelementsofadvancedvalvecontrol strategies, including the impact of refrigerant migrationcontrol strategies. The experimental results confirm that 30% to50% improvements in cyclic coefficient of performance (COP)are possible using improved expansion val
4、ve controls, while thebenefits of preventing refrigerant migration do not outweigh theadditional cooling achieved if refrigerant continues to flowthrough the expansion valve during the compressor OFF period.INTRODUCTIONNearly40%oftotalU.S.energyconsumptionin2011wasconsumedinresidentialandcommercialb
5、uildings,asfoundinthe U.S. Building Energy Data Book (EIA 2012). Supermar-kets are one of the most energy-intensive types of commercialbuildings, as significant electrical energy is used to maintainchilled and frozen food in both product display cases andwalk-in storage coolers. A typical supermarke
6、t consumesroughly 2 million kWh (6824 MMBtu) annually, and roughlyhalf is for refrigeration (Zhang 2006). Therefore, improvingthe energy efficiency of the supermarket refrigeration systemdirectly affects the stores operating expense.Forresidential,commercial,andindustrialair-conditioningand refriger
7、ation systems, vapor compression is the mostcommon technology. The vapor-compression cycle uses acirculating refrigerant as the medium that absorbs andremoves energy from the cooled space and subsequentlyrejects that energy elsewhere. An ideal vapor-compressioncycle and its pressure-enthalpy relatio
8、n are shown in Figure 1.Evaporator superheat, which is defined as the differential ofevaporator outlet temperature and the saturation temperatureofevaporatorpressure,isaquantitythathasanimportantrela-tion to system efficiency. Optimal efficiency is generallyobtained with a few degrees of evaporator
9、superheat.However, maintaining a nonzero value of superheat is criticalto prevent damaging the compressor with two-phase flowrefrigerant. The primary method of effectively regulating thesuperheat is modulating the expansion valve opening withmechanical or electrical feedback control mechanisms.Inthe
10、HVACChenetal.2002).EEVstypicallyregulatesuperheatbasedonstandard proportional-integral-derivative (PID) controllers.The improved control afforded by EEVs has the potential toavoid valve hunting, but poorly tuned EEV controllers maystill result in undesirable system behavior. Micro-electro-Quantifyin
11、g Efficiency Gains ofRefrigeration Systems UsingAdvanced Expansion Valve TechnologyKaimi Gao Bryan P. Rasmussen, PhD, PEStudent Member ASHRAE Member ASHRAEKaimi Gao is a doctoral candidate and Bryan P. Rasmussen is an associate professor of Mechanical Engineering in the Department ofMechanical Engin
12、eering at Texas A Aprea and Mastrullo 2002;Lazzarin and Noro 2008), selecting the parameters of EEVcontrollers can prove challenging. System and valve nonlin-earities can be significant, and several researchers havereported the need to schedule the controller gains based onthe operating conditions t
13、o avoid poor performance(Outtagarts et al. 1997; Doyle 2000). Multievaporatorsystems present additional difficulties due to the couplingbetween evaporators.This paper proposes using a cascaded control method-ology, capable of inherently addressing each of these chal-lengeswhileretainingthesimplicity
14、ofPID-typecontrollers.Previous studies have documented the benefits of this archi-tecture in providing nonlinear compensation (Elliott andRasmussen 2010) and in helping to decouple the dynamicsbetween multiple evaporators (Elliott et al. 2011).The cascaded controller has two feedback control loops.T
15、he first control loop regulates evaporator pressure with aproportional controller. Using high gains for this innercontrol loop provides effective nonlinear compensation butrequires an expansion valve with fast actuation. The secondcontrol loop uses a conventional proportional-integral (PI)controller
16、 to regulate evaporator superheat. The schematicstructureofthecascadedcontrollerisshowninFigure3.Theinner and outer control loops exploit the inherent time scaleseparation between evaporator pressure and superheatresponses. The system can compensate for sudden changesin compressor speed, as these dy
17、namics have a dominanteffect on evaporator pressure. In contrast, changes in ambi-ent air conditions result in slower drifts in evaporator refrig-erant outlet temperature and can easily be regulated by theouter control loop.Figure4showstheperformanceofthedesignedcascadedcontroller with MSEVs on a mu
18、lti-evaporator refrigerantsystem. The system was given superheat step changes. TheMSEVs reached the superheat setpoint in 220 seconds afterstart-up. When the superheat setpoint changed, the MSEVsrespondedquicklytotrackthesuperheatsetpoint,withlimitedcoupling evident between the evaporators. This dem
19、onstratesthe cascaded controllers quick response time and ability toreach operating conditions.Figure 3 Schematic structure of cascaded controller.Figure 4 Superheat setpoint change demonstrating control loopperformance.Published in ASHRAE Transactions, Volume 122, Part 2 92 ASHRAE TransactionsEXPER
20、IMENTAL TEST SYSTEMThe experimental system used in this efficiency study isa multievaporator commercial supermarket refrigerationsystem. A photograph and the schematic of the system areshown in Figure 5. These systems typically operate 24 hoursa day with periodic compressor ON-OFF cycles and with tw
21、o ormore defrost cycles each day. The compressor ON-OFF cycle isautomatically triggered by the discharge air temperature ther-mostat installed in the display case. In this experimentalsystem, the unit is designed to shut off when the discharge airtemperature falls below 1.1C (30F). The system has th
22、reeevaporators retrofitted with variable-speed fan control and asingle water-cooled condenser.To compare system performance using different controlstrategies, MSEVs are installed in parallel with the factory-installed TXVs. Additionally, two control algorithms areapplied with the MSEVs: 1) the propo
23、sed cascaded controllerand 2) the default first-generation adaptive PID control(APID) released in early 2014 and provided by the valvemanufacturer. The gains of the cascaded controller are tunedexperimentally to achieve acceptable superheat regulation,andthegainsoftheAPIDaretunedaspermanufacturergui
24、de-lines.Finally, two additional strategies are tested to determinethe impact of refrigeration migration management on thecyclicefficiencyofthesystem.Thisrequirestheinstallationofbypassvalvestoensurethatrefrigerantchargecanbevirtuallyisolatedduringperiodswherethecompressorisnotoperating.EXPERIMENTAL
25、 METHODOLOGYTo evaluate the impact of expansion valve control strate-gies on the transient efficiency of these refrigeration systems,two compressor cycling scenarios are proposed. In the firstscenario, the compressor cycling is controlled by the displaycase thermostat. Depending on the expansion val
26、ve actions,the time required to lower the display case temperature to therequired setpoint will be different, resulting in different cycletimes. For this case, the system is allowed to cycle until arepeatable pattern is observed, and then the efficiency is aver-aged over five full cycles. In the sec
27、ond scenario, the cycletimeisfixed,either5minon/5minoff,or5minon/20minoff.This allows a comparison based on the differences in start-uptimesandaveragecoolingcapacitiesduringequalcycletimes.Additional tests are also performed to compare the impact ofrefrigeration migration strategies as well as the p
28、erformanceof the system at different ambient temperatures.Figure 5 (a) Experimental system and (b) schematic of experimental system.Published in ASHRAE Transactions, Volume 122, Part 2 ASHRAE Transactions 93The system is equipped with a full suite of temperature,pressure, and refrigerant flow sensor
29、s, as well as a powersensortodeterminethetotalelectricalpowerconsumedbythesystem compressor and fans. The uncertainties of the majortransducers are listed in Table 1. The cooling capacity of thesystem is calculated using both refrigerant-side and air-sidemeasurements. Equation 1 defines the cooling
30、on the refrig-erant side:(1)where hinletand houtletare the refrigerant enthalpies at theevaporator inlet and outlet, respectively. These values arecalculated based on measured pressure and temperature atappropriate locations on the system. Refrigerant mass flowrate is measured at the expansion valve
31、 inlet.Equation 2 defines the cooling on the air side of the evap-orator:(2)Noting that the humidity inside the closed display case is low,the cooling capacity is calculated using the specific capacityof dry air cp. The flow rate of air varies as a function offan speed and is measured prior to the t
32、ests using a calibratedairflowmeter.Tair,outandTair,inaretemperaturemeasurementsfrom the thermocouples installed at the fan inlet and the evap-orator outlet. The reader should note that the refrigerant-sidecooling capacity measurement is only valid when thecompressor is running. When the compressor
33、is shut down,refrigerant may continue to flow through the expansion valve,but this measured flow rate is not uniform throughout theevaporator and is not representative. Thus, the air-sidemeasurement is preferred during the compressor OFF cycle.The average cooling per cycle is defined by Equation 3:(
34、3)One cycle is defined as the period of time for the compressorto turn on and turn off once, Tcycle. Q() is the instantaneouscooling as a function of time. Qcycle(t) is the average coolinggained by the system over a single cycle.Similarly, the cyclic coefficient of performance (COP) ofthe system is
35、defined by Equation 4:(4)whereW()istheinstantaneousworkinputtothesystem.Thisdefinition of COP captures the residual cooling achieved afterthe compressor is shut off as well as the transient coolingachieved during start-up.EXPERIMENTAL RESULTS:THERMOSTAT-CONTROLLED CYCLINGIn this first scenario, the
36、system was operated for a mini-mum of two hours to achieve five repeatable ON-OFF cycles,where the compressor cycling was triggered by the displaycase thermostat. These tests were intended to benchmarksystem efficiency under standard operating conditions. Thedischarge air temperature setpoint was 1.
37、1C (30F). For alltests, the superheat setpoint was 8C (14F) above saturation,whichisthefactorydefaultsuperheatlevelfortheTXVsunderthese particular operating conditions.Performance Comparison ofAdvanced Expansion Valves and ControllersThe first set of experiments compared the system perfor-mance with
38、 TXVs versus the advanced expansion valves andcontrollersproposedinthispaper.Twoadvancedvalvecontrolconfigurations were tested, the MSEV with the manufacturerdefault APID and the MSEV with a cascaded controller(MSEV+CC), as shown in Figure 6.The performance data listed in Table 2 are the average off
39、ivecompressor ON-OFF cyclesafterthesystemhadconvergedto a repeatable cyclic behavior. Significant efficiency gains inthe cyclic COP were observed with the use of the MSEV witha default controller (+35%) as compared to the TXVs. This isprimarily due to the faster actuation capability of the MSEV,whic
40、h results in faster convergence to steady-state pressureconditions.AlthoughtheaveragesuperheatoftheMSEVwiththe default controller is higher than the TXV, some coolingwas gained in the OFF cycles, which resulted in a total coolingincrease.TheTXVeffectivelyregulatessuperheat,butdoessoattheexpenseofpul
41、lingdownpressureataslowerrate,result-ing in reduced cooling capacity during start-up. The cascadedcontroller achieves even higher efficiency (+53%), as itexploits the fast actuation capabilities further in quickly regu-lating both evaporator pressure and superheat.Low-Temperature and Medium-Temperat
42、ureComparisonRefrigerated food products, such as fresh meats andproduce, require strict environmental conditions, withtemperatures that range depending on the particular item(from 35C to 20C 31F to 68F). Effective and efficientTable 1. Sensor UncertaintyType T Thermocouples 0.5C (0.9F)Pressure Trans
43、ducer (Evaporator) 21 kPa (3 psi)Pressure Transducer (Condenser) 21 kPa (3 psi)Flowmeters (9) 0.1 L/min (0.26 gal/min)Power Transducer 0.5% ()Qrefmrefhoutlethinlet=QaircpmairTair outTair in=mairQcycletQttTcycle+dTcycle-=COPcycletQttTcycle+dWttTcycle+d-=Published in ASHRAE Transactions, Volume 122, P
44、art 2 94 ASHRAE Transactionscontrol of supermarket refrigeration systems over a widerange of conditions is essential. Other vapor-compressionsystem applications (residential air conditioning, laboratoryor greenhouse conditioning, data center cooling, etc.) oftenfacesimilardemandsforeffectiveoperatio
45、noverawiderangeof ambient conditions.A distinct advantage of the proposed cascaded controlalgorithm is the inherent nonlinear compensation it provides.This characteristic has been demonstrated for various HVACcomponents (Price et al. 2015), and for expansion valves inparticular (Elliott et al. 2011)
46、. In this section, the performanceof the cascaded controller and the manufacturers defaultAPID controller are compared in low- and medium-tempera-ture conditions. In the low-temperature case, the system ther-mostat turns the compressor off at 1.1C (30F) and turns thecompressor on at 0C (32F). For th
47、e medium-temperaturecase, the thermostat settings are 11C and 16C (52F and61F), respectively. As before, the superheat setpoint is 8C(14F). Both controllers were tuned specifically for the low-temperature condition.Table 2. Performance Comparison for Thermostat Controlled CyclingExpansion Valve and
48、ControllerCooling,kWCooling,tonsPower,kWPower,tonsCOPCOPImprovementThermostatic expansion valve (TXV) 26.2 7.4 15.1 4.3 1.7 MSEV with default controller (MSEV+APID) 34.9 9.9 15.3 4.4 2.3 35%MSEV with cascade controller (MSEV+CC) 41.2 11.7 15.8 4.5 2.6 53%Figure 6 Thermostat-controlled cycling comparison: (a) average superheat, (b) system cooling, and (c) power consumption.Published in ASHRAE Transactions, Volume 122, Part 2 ASHRAE Tr
